Projects in LVVD involve the use of infectious DNAs to derive selected mutant RNA viruses. The term "infectious DNA'" refers to a full-length copy of the genome of an RNA virus such that RNA transcribed from the DNA in vitro is infectious when transfected into tissue culture cells. This technology is possible for all known positive-strand RNA viruses, because genomic RNA for any of this group is infectious. Since mutagenesis of DNA is technically simpler than direct mutagenesis of RNA and techniques are more site-specific, it is easiest to mutagenize DNA and then to generate RNA for transfection, in order to create new mutant viruses. LVVD works on the vector-borne viruses dengue (DEN), Japanese encephalitis (JE), and West Nile (WN), in keeping with our major regulatory mission. Using recombinant infectious DNA technology, we have recently concentrated our efforts on a determination of the role of a conserved "stem and loop" secondary structure (the 3'SL), located in the 3'noncoding region of all flavivirus RNAs, in virus replication. The 3'SL, but not its nucleotide sequence, is conserved in all flavivirus genomes. By making mutant DEN viruses with chimeric DEN/WN 3'SL nucleotide sequences and conversely by making mutant WN viruses with chimeric WN/DEN 3'SL nt sequences, we have demonstrated that bulges created by the apposition of unpaired nucleotides in the long stem of the 3'SL are crucial for replication and host range of both DEN virus and WN virus. Also out of this work, LVVD has developed a promising candidate dengue vaccine virus for serotype 1 among the 4 serotypes, and evidence suggests that the same set of mutations attenuates the remaining 3 serotypes as well. Currently, we are testing candidate 3'SL mutant WN viruses that also were derived in this project to determine whether they are attenuated in a mouse model of WN neurovirulence. LVVD conducts other research relative to the mechanism of replication of flaviviruses based on the infectious DNA technology. LVVD has an active program related to bioterrorism. We are studying the feasibilty of administering live vaccine in the form of RNA, using an existing but unlicensed live Venezuelan Equine Encephalitis virus vaccine as a model. VEE is an alphavirus, and alphaviruses, like flaviviruses, are positive-strand RNA viruses. Hence the RNA is infectious in tissue culture cells. If a VEE vaccine could be administered as RNA, this would speed up the manufacturing process, obviate any problems related to adventitious agents in tissue culture, and provide for a more rapid response to a bioterrorist attack with one of these agents. LVVD is also involved in development of an ELISA to test potency of rabies vaccines. Currently, this potency assay requires hundreds of mice, is very expensive, not very repeatable, and undesirable due to concerns for the humane treatment of experimental animals. Methods for evaluating attenuated vaccine candidate viruses. No new work was done on this project during 2002-2003. However, future work is planned. The dengue virus (DEN) protease is a heterodimer of nonstructural protein 2B (NS2B) and NS3. Previously, we performed a mutagenesis study of some of the residues in NS3 of DEN type 2 (DEN2) predicted to be involved in substrate binding. In all, 46 mutations were analyzed for their effect on cleavage in vitro. Thirteen of these mutants with wild-type or nearly wild-type activity were analyzed by reverse engineering into a full-length DEN2 cDNA clone and testing mutant RNA transcripts for infectivity: 6 were non-infectious, while 7 were recovered as virus. The recovered viruses had wild-type growth kinetics. Now, 8 members of a previously-made set of mutations in the region of DEN4 NS2B essential for protease activity have been engineered into an infectious DEN4 clone made by Robin Levis. Virus was recovered from 5 of these, while 3 were lethal. One of the recovered viruses was a partial revertant. Several of the viruses had a second-site mutation at another location in NS2B, changing a specific met residue to ile or val, complicating the analysis of the effect of the introduced mutations on the phenotype. One of the recovered viruses had a significant growth defect. We had previously observed that transcripts made from DEN2 cDNA clones with short deletions at the 3' end are infectious. Constructs missing up to 6 nt are viable, while deletion of 8 or more nt is lethal. Deletions of 7 nt are viable or lethal, depending on the presence and composition of a restriction site overhang at the 3' end of the run-off transcript. In all but one case, the 3' end sequences of the recovered viruses are wild-type; virus recovered from the 7 nt deletion has a 3' end sequence that differs from wild-type at 1.5 positions, but this mutant virus appears to grow with wild-type kinetics. To investigate this further, we have continued our collaboration with the Padmanabhan lab, now located at Georgetown. Several small 3' end deletions are being analyzed in their in vitro RNA replication system. So far, it is clear that the deleted RNAs are replicating, and analysis of the 3' end sequences of replicating RNAs is ongoing. We have two other ongoing collaborations with the Padmanabhan group. (1) Several double mutants in DEN2 NS3 were constructed, and are being expressed in E coli and purified, for subsequent crystallographic analysis. We are also trying to express and purify a dimer between NS3 and the protease domain of NS2B. (2) We have attempted to make a DEN2 replicon expressing green fluorescent protein (GFP). The initial results do not look promising, in that the level of GFP expression is very poor. Work is underway to change transgenes. Applications of infectious cDNA technology to RNA virus vaccine development. In a collaboration with Steve Whitehead and Brian Murphy at the NIH we had introduced a 30 nt deletion - which they had shown was attenuating for a DEN4 infectious clone - in the 3' non-coding region of our DEN1 WP infectious clone. The recovered DEN1 mutant virus grew more slowly than its parent, and made a smaller plaque. This virus has now been shown to be immunogenic and attenuated in monkeys. These results have been submitted for publication to the Journal of Virology. In collaboration with E Kelly, we had sequenced and made an infectious clone of a DEN2 PDK50 vaccine candidate and its virulent parent. Clone-derived viruses were comparable to their parents for growth kinetics in tissue culture cells. The cloned DEN2 PDK50 was manufactured under GMP by K. Eckels and his collaborators, and this virus was tested in monkeys for attenuation and immunogenicity; this study should be completed soon. We had also made an infectious clone of the Army's candidate live attenuated Japanese encephalitis virus (JEV) vaccine, a vero cell adapted version of a PDK cell derivative of the Chinese JEV live vaccine strain SA14 14-2. Clone-derived virus grew with the same kinetics as its parent in cell cultures. This virus was tested in mice and has been shown to be immunogenic. Finally, with help from Bangti Zhao I have been trying to make an infectious clone of DEN3 in the pRS424 yeast-E. coli shuttle vector. I was able to make a full-length DEN3 clone in yeast, but all attempts to recover this clone in E coli failed; the only plasmids which can be recovered have suffered large deletions of the DEN3 sequences. Subsequently, I introduced the mutF mutation into the 3' end of the DEN3 clone in yeast. Both the WT and mutF yeast clones were used in PCR to amplify the full 10.7 kb genome, using Expand polymerase. Full-length run-off RNA transcripts were made from these PCR products. Electroporation of these transcripts into cells resulted in the recovery of WT and mutF DEN3 viruses. Partial sequence analysis has shown that the mutF mutation has been retained, but that these viruses are not identical in the structural gene region, differing at two amino acids out of the approximately 1 kb sequenced. Preliminary analysis of the phenotype shows that DEN3 mutF makes a very small plaque, is severely growth restricted on LLCMK2 cells, and does not grow at all on mosquito C6/36 cells. This result is complicated to interpret, since there are an as yet unknown number of adventitious differences between the WT and mutF viruses, due to the use of full-length PCR in their derivation. Plans are underway to sequence both viral genomes, after repeating the phenotypic analysis. Also, work is underway to try to stabilize the DEN3 clone in E coli: I plan to introduce insertions at various locations in the left half of the genome; based on previous work in DEN2, I expect this to allow the clone to grow in E. coli. Subsequently, these insertions can be removed by restriction digestion prior to RNA transcription to make infectious RNA. Molecular approaches to enhance safety of RNA virus vaccines. Flavivirus genomic RNAs contain a conserved stem-loop structure within the 3'-noncoding region (3'-SL). One cDNA-derived DEN2 virus that contained a 7-bp substitution of WN genomic 3'-SL nt sequences for the analogous nt sequences of wt DEN2 3'-SL was severely restricted for growth in mosquito cells but replicated like wt D2 virus in monkey kidney cells. Subsequently, the deletion and substitution mutations that differentiated DEN2mutF with respect to the wt D2NGC genome were introduced into an infectious full-length DNA derived from the genome of DEN1 strain WestPac virus. This "DEN1mutF" virus grew to wt titers in LLCMK2 cells but did not replicate in cultured mosquito cells, recapitulating the phenotype of the original DEN2mutF virus. DEN1mutF virus had all properties of a suitable vaccine candidate; it was highly immunogenic, inducing similar titers of neutralizing antibodies in sera compared to wt, and severely impaired in its ability to cause viremia. The DEN1mutF virus is of particular interest to vaccine developers, since DEN1 virus is the most common cause of serious dengue illness in SE Asia. Monkeys immunized with a single dose of DEN1mutF virus were resistant to challenge with wt DEN1 virus more than one year post-immunization. A revertant of DEN1mutF virus was also created by serial passage in mosquito cells. After three blind passages of 8 days each, or a total of 24 days, a fully reverted DEN1mutF virus was detected, DEN1mutFRev. The complete genomes of DEN1mutF and DEN1mutFR viruses were sequenced and compared to that of the wt parent D1 virus. D1mutFR RNA contained three additional mutations vs D1mutF, in NS1, NS5, and in the 3'-SL. Any or all of these mutations could affect RNA synthesis, the presumptive defect in mutF viruses when grown in mosquito cells, based on earlier data. A set of DEN1mutant virus genomes was constructed, containing all possible combinations of the 3 mutations detected in the DEN1mutFR genome compared to that of DEN1mutF. This year, we showed that the point mutation in the 3'SL plus either the mutation in NS5 OR the mutation in NS3 were together required for the phenotypic reversion of DEN1mutFR virus. (That is, a virus with only the NS3 and NS5 point mutations but lacking the 3'SL mutation that was present in the DEN1mutFR genome was NOT fully revertant.) We also showed that DEN1mutFR virus was as attenuated in rhesus monkeys as was its parent DEN1mutF virus. In other words, the host-range restricted phenotype of DEN1mutF virus was independent of its attenuation phenotype in monkeys. Scientists in LVVD created DEN3 and DEN4 mutant viruses containing the mutant F set of mutations, and these viruses were found to be host range-restricted mutants, like DEN2mutF virus and DEN1mutF virus. The level of attenuation of the former viruses in animals has yet to be determined, but this finding shows that it may be possible to construct a tetravalent mutant F dengue vaccine. West Nile virus is a known cause of fever/arthralgia/rash syndrome and encephalitis in N Africa and the middle East. However, no known case of WNV infection of humans had occurred in N America until the past few years, when isolated cases of WN encephalitis have been documented in NYC, NJ and MD. This year, clinically evident infection with WNV spread throughout the eastern two-thirds of the US, with a dramatic increase in deaths. We obtained a full-length DNA copy of the WNV strain Wengler genome from Dr. Vladimir Yamshchikov at the University of Virginia. Dr. Li Yu created a large number of mutations in the nt sequence of the 3'SL in WNV genomic DNA and evaluated the growth phenotypes of the WN mutant viruses that were thus derived, in LLCMK2, BHK, vero, and C6/36 cells. Findings in brief are as follows: For WN, substitution of the "bottom" half of the WN 3'-SL by DEN2 nt sequences resulted in a viable mutant virus that replicated efficiently in both monkey and mosquito cells. This was in contrast to results previously obtained in the context of the D2 genome. A second WN mutant virus, containing a substitution of D2 nt sequences for the wt WN nt sequences in the "top" half of the WN 3'-SL was non-viable. A revertant of this virus displayed a host range restricted phenotype in mosquito cells, a la DEN2mutF virus, and its genome contained spontaneous mutations in the region of the 3'-SL involved in mutF. Comparison of the results for WNV with those earlier obtained with DEN2 V suggests that conserved bulges in the long stem of the 3'-SL, at the boundary of the top and bottom portions previously defined, are required for virus replication. Introduction of a novel bulge into the long stem of the WN 3'SL, by substitution of nts in that segment with analogous DEN2 2'SL nts, resulted in a mutant WN virus that replicated with 100-fold greater efficiency in C6/36 cells compared to wt WN parent virus. Taken together with the known properties of DEN mutant F viruses (see above), the data suggested that the bulge (in the lower-most portion of the long stem) is an enhancer of replication competence that is specific for mosquito cells. Mouse neurovirulence studies are currently underway to determine whether any of the 3'-SL mutant viruses have potential merit as vaccine candidates. Methods to enhance efficiency of RNA virus vaccine manufacture. NOTE: At the close of the 2002 academic and fiscal year, Dr. Robin Levis accepted a position as Regulatory Coordinator for the Division of Viral Products and therefore resigned her position as Senior Staff Fellow in LVVD. The work in this project was formerly under her supervision. Thus no new work was done on this project in 2003. However, it is possible that the lab will continue with these projects at a future time. We have previously shown that LLC-MK2 cells expressing the first nonstructural protein in the dengue 2 genome, NS1, can complement viruses which contain lethal mutations in the NS1 gene. We have identified a 25 amino acid region at the carboxy-terminus of the NS1 gene which is not able to be complemented by NS1. This region must be involved in an essential cis-function of the NS1 protein such as post-translational processing of the polyprotein. More detailed deletion analysis of this region has shown that 10 amino acids located -13 to -23 from the carboxy-terminus of the NS1 protein are involved in this cis-function of the NS1 protein. In studies designed to understand the transport of NS1 in infected cells, we have demonstrated that the NS1 protein associates with viral structural proteins, prM and E in infected mammalian cells (the cell line LLC-MK2). A similar analysis of NS1 in virus containing supernatants of infected cells did not show association of NS1 with viral structural proteins. This suggests that the association between NS1 and the viral structural proteins in the cell cytoplasm of mammalian cells may be important for viral protein transport and targeting of viral proteins to the cell membrane where virus maturation occurs. In contrast, similar studies done in a mosquito cell line failed to demonstrate an association between NS1 and the viral structural proteins. As for the mammalian cell studies, NS1 was not associated with the viral structural proteins in virus-containing cell supernatants. Studies could be done to determine the nature of the interaction between NS1 and the structural proteins and what this interaction means in the virus life cycle.